What project or projects do you maintain and what was your motivation for creating those projects and releasing them as open source software?
Not everything I’ve worked on over my career has been open source, but most of it has been. I have 100s of open source projects on GitHub and npm. I am able to create lots of value in the world when I don’t restrict how people can use my work. Publishing things under an open source license lets anyone use my code however they like.
For example, my business Study Notes is doing quite well at making money even though the code is available on GitHub. Part of my motivation is to prove that it’s possible to make a profit and share your code with the world :)
If you created any of those projects, were they meant to solve a specific problem you faced, or were they born out of a larger opportunity you saw?
Out of all my projects, Standard is the one that was most clearly designed to solve a specific need that I had at a very specific point in time.
While working on WebTorrent, I kept getting useful code contributions that had minor style or functionality issues, e.g. not matching the existing code style, forgetting to handle errors, or creating paths in a way that doesn’t work cross-platform (“/” vs. “\”). Whenever this happens, a maintainer can do one of two things: 1) provide feedback and hope that it gets fixed (anecdotally 50% of the time, it doesn’t because the contributor is too busy, or just goes silent) or 2) accept the code, but immediately make a followup commit to fix the issues.
I realized that there’s actually a third, better way. With a linter, you can catch code style issues and bugs before a code contribution comes in. That was the motivation for Standard. Catch style errors before they’re submitted in PRs to save precious code review time by eliminating back-and-forth between maintainer and contributor.
But because the WebTorrent project is split across dozens of GitHub
repositories, and I didn’t want to paste in a multi-hundred line linter
configuration in each project, I created a module (
standard) that all
repositories could depend on. This made Standard require no configuration, which
was unique among linting tools. It was really easy to enforce consistent style
across all my GitHub projects.
When I first started Standard, I was going to call it webtorrent-style, or feross-style, since it was literally my own stylistic preferences (plus some non-controversial error catching rules). But I thought it would be amusing to pick a very opinionated name, so I chose “standard”, which was available on npm at the time.
Since then, hundreds of thousands of people have adopted Standard. Most are just happy to have a sane default style guide that takes a few seconds to set up. Some use it as a way to end the constant bickering over code style that takes place on their teams.
How has the project evolved since you first got involved or first released it?
Since the beginning, Standard has been adding new rules as ESLint releases them. The most important rules are designed to catch errors. But we also add rules to reduce inconsistent code, i.e. if there are two ways to do the same thing, we prefer to only allow one way. This makes writing code fater – as you don’t have to consider all the different ways you could do something. And there will only ever be the one way across the whole codebase. Standard is all about consistency.
standard command line program has also gotten a lot faster to run. We use
aggressive caching and only run a single linter (ESLint, instead of
ESLint+JSHint+JSCS, like we did at the beginning). We also use ESLint’s
--fix feature to provide really great automatic fixing of
non-standard code. Just run
How do you spend your time on those projects? (i.e. Developing, managing the community, triaging issues, etc.)
I spend the most time triaging issues and considering new rules.
There is a really delicate balance between enforcing new rules that I would like to see adopted by the community versus being sensitive to how much work it is to update your code to the new rules. Before I consider any rule change or addition, I run a very thorough test suite that runs Standard on around 400 community repos to see what the effects of the rule change would be on the community. Evolving towards stricter rules, especially ones that enforce the use of new language features from ES2015 and beyond, needs to be done carefully and respectfully.
How would you describe the community around projects you participate in? What are your favorite and least favorite aspects?
Standard is maintained by a dedicated group of contributors who work tirelessly to help triage issues, fix bugs, and debate the pros/cons of new rules.
Standard is an OPEN Open Source project, so whenever someone makes a significant contribution to Standard, they become an owner of the project, on both GitHub and npm. As of today, there are 16 contributors in addition to myself. Many of them are popular package authors in the community: flet, jprichardson, dcposch, mafintosh, dcousens, othiym23, maxogden, watson, jb55, bret, linusu, timoxley, xjamundx, rstacruz, reggi, yoshuawuyts.
Standard is a true community effort, and I couldn’t do it without all the help I get from these nice folks <3
What keeps you involved in those projects? Do you have long term plans for maintaining your involvement?
I stay involved in Standard because I feel like it’s really valuable to the community.
“I deal with all the code style arguments so your team doesn’t have to.”
Instead of 1000s of teams having the same debates over every ESLint rule, we have them once in the Standard repo and then we can all move on. It’s kind of ironic that I started Standard so I wouldn’t have to talk about style anymore, and now I do that a whole lot more. But since it’s for a good cause, I’m willing to spend my time this way to help out the community that I love.
What is the most important thing someone submitting an issue or patch should know?
My inbox is overflowing with issues. Most open source maintainers inboxes are the same way.
Please include as much information as possible, like the version of the package you’re using, the Node.js version, the OS version, as well as detailed steps to reproduce the issue.
Also, don’t forget to be nice! Almost all open source maintainers are unpaid volunteers. We’re doing this because we love it, but dealing with thousands of issues over the months and years can really take a toll. If you’re mean, you’re less likely to get the help you seek, but worst of all, you contribute to maintainer burnout. Burnout is a huge and underreported problem in open source.
What’s your development environment right now?
As Dominic Tarr says, “If you want to write fast software, use a slow computer”.
What was your first development environment? Do you miss anything from it?
I used to make sites with PHP. One of the things I miss is the ease of deployment. Just FTP your code up to the server and your site is live! Nothing to configure! You can even edit your PHP source files live on the server, although there are risks associated with this.
People give PHP a lot of crap, but it actually deserves a lot of credit for the things is gets right.
Where do you see the open source software community headed?
I don’t claim to be able to foresee where the community is headed. But I do hope that we address the open source funding problem. As mentioned before, most open source maintainers are unpaid volunteers, even though our software is frequently used by Fortune 500 companies, sometimes in very critical contexts (security, uptime, etc.).
The OpenSSL project was run by a handful of unpaid volunteers, even though it powers most of the internet’s secure communications and commerce. Recently, the team got funding from a few big companies and now the lead developer gets paid full-time. But the other volunteers are still working for free.
We need a better model, because the current one doesn’t feel sustainable to me.